M1TUS Analysis
Benjamin Kop
01/12/2022
Storage of analysis code and output
Experiment I
Comparison to baseline
We first test whether TUS at parameters used in prior research reduced MEP amplitude from baseline, and whether that is also the case for stimulation of a control region (active control) or administration of a sound alone (sound-sham). The relevant conditions here are: 500 ms of TUS/AC at 32.5 W/cm2 without a masking stimulus, sound-sham, and baseline (i.e., TMS-only)
| Baseline |
| 500ms_AC |
| 500ms_TUS |
| 500ms_Mask |
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.1123650 0.11347671 11.00082 9.802584 9.014327e-07
## Condition500ms_AC -0.1194853 0.05216414 11.01509 -2.290563 4.270491e-02
## Condition500ms_Mask -0.1437009 0.04522184 11.02462 -3.177688 8.774258e-03
## Condition500ms_TUS -0.1398309 0.06269135 11.08148 -2.230465 4.731814e-02## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Condition | 0.48 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Stimulus duration (direct comparison)
Now we test whether there is an effect of stimulus duration on MEP amplitude, which inherently also involves a direct comparison between on-target, active control, and sham conditions.
| TUS |
| AC |
| Sham |
| short |
| long |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## procedure 0.29978 0.14989 2 11.048 1.2998 0.311198
## stimulusDuration 1.16168 1.16168 1 11.024 10.0741 0.008833 **
## procedure:stimulusDuration 0.14904 0.07452 2 11.009 0.6462 0.542784
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## ----------------------------------------------------------
## procedure | 0.19 | [0.00, 1.00]
## stimulusDuration | 0.48 | [0.10, 1.00]
## procedure:stimulusDuration | 0.11 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Stimulation intensity (free-water)
Now we want to determine whether there is any difference in MEP amplitude between trials where stimulation was delivered at 32.5 W/cm2 (originally 10 W/cm2) and 65 W/cm2 (originally 20 W/cm2). These intensities were applied both to the hand area and the contralateral face area.
| TUS |
| AC |
| low |
| high |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## procedure 0.070792 0.070792 1 11.039 0.6868 0.4248
## intensity 0.082657 0.082657 1 11.110 0.8020 0.3895
## procedure:intensity 0.010329 0.010329 1 11.012 0.1002 0.7575## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## ---------------------------------------------------
## procedure | 0.06 | [0.00, 1.00]
## intensity | 0.07 | [0.00, 1.00]
## procedure:intensity | 9.02e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Simulated intensity (TO BE COMPLETED)
To investigate whether there is a dose-response relationship between the estimated intensity intracranially and modulation of motor cortical excitability, we test for a relationship between simulated intensity and change in MEP amplitude from baseline (and from active control and sham conditions)
Work in progress.
Masking
We want to determine whether audible differences between stimulation sites (i.e., left-hemispheric hand motor area and right-hemispheric face motor area) may confound our results, which would be illustrated by an interaction between procedure(TUS/AC) and masking (unmasked/masked)
| TUS |
| AC |
| unmasked |
| masked |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## procedure 0.0163725 0.0163725 1 11.021 0.1535 0.7027
## masking 0.0011414 0.0011414 1 11.080 0.0107 0.9195
## procedure:masking 0.0001484 0.0001484 1 11.038 0.0014 0.9709## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -------------------------------------------------
## procedure | 0.01 | [0.00, 1.00]
## masking | 9.65e-04 | [0.00, 1.00]
## procedure:masking | 1.26e-04 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Preperatory learning
It is possible that expectation of the TMS pulse underlies the observed motor inhibition for 500 ms ultrasonic and auditory stimuli. In this scenario, we would expect that during the first block of 500 ms, participants will show a decrease in MEP amplitude throughout the block as they learn to predict TMS by the auditory stimulus. During this experiment, participants receive two different stimulus durations, in addition to the sound-sham condition having slightly altered timing (i.e., onset ~50 ms earlier). Therefore, it would also be reasonable to expect that in each subsequent block, a number of trials will be needed to re-adjust to the relevant stimulus duration. This number of trials needed would be much lower than the amount of trials needed for initial learning in block 1. Notably, this can be confounded with a 15 second rest period prior to each block, and the known phenomenon that the first TMS pulse typically evokes a higher amplitude MEP.
First longer stimulus duration block for each participant: 
## [1] "For trial per block 1-20"## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 0.20519 0.20519 1 11 2.2296 0.1635
## [1] "For trial per block 1-10"## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 0.87989 0.87989 1 11 8.2959 0.01496 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Subsequent longer stimulus duration blocks for each participant:
## [1] "Trial number per block:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 0.12656 0.12656 1 11.085 1.0743 0.3221
## [1] "Trial number in total:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.033639 0.033639 1 10.852 0.2905 0.6008
Experiment II
Comparison to baseline
As in Experiment I, we first test whether TUS at parameters used in prior research reduced MEP amplitude from baseline, and whether that is also the case for stimulation of a control region (active control) or administration of a sound alone (sound-sham). The relevant conditions here are: 500 ms of TUS/AC at 6.35 W/cm2 without a masking stimulus, sound-sham, and baseline (i.e., TMS-only). Here, we also include TUS administered with a masking stimulus a pre-registered.
| Baseline |
| Sham |
| TUS_10W_M |
| AC_10W_NM |
| TUS_10W_NM |
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.9443794 0.04668147 26.02677 20.230282 1.916112e-17
## ConditionAC_10W_NM -0.2216959 0.03964922 26.10652 -5.591431 7.009302e-06
## ConditionSham -0.2232467 0.04161195 26.01801 -5.364966 1.283150e-05
## ConditionTUS_10W_M -0.2090008 0.04143463 26.15454 -5.044109 2.944288e-05
## ConditionTUS_10W_NM -0.1815641 0.03776016 26.24006 -4.808352 5.453072e-05## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Condition | 0.40 | [0.19, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Stimulation intensity & masking (direct comparison)
For Experiment II we have a more balanced experimental design, making it possible to include both intensity and masking in a single model. Here, we address the following questions:
1. Is there a difference in MEP amplitude following on-target or active control stimulation? (direct comparison)
2. Is there a difference in MEP amplitude when higher free-water intensities are administered?
3. Do audible differences possibly underly observed effects (interactions with masking)
A three-way interaction between stimulation site (on-target/active control), intensity (6.35/19.05 W/cm2), and masking (unmasked/masked) is relevant. For example, if audible differences play a role, we would expect and interaction between stimulation site and masking. If intensity is related to motor inhibition, we would expect an interaction between stimulation site and intensity, where higher intensities only result in further MEP attenuation for on-target, but not active control stimulation. Intensity and masking may interact with eachother, because higher intensity stimulation is louder, and thus masking may be less efficacious.
| TUS |
| AC |
| unmasked |
| masked |
| low |
| high |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## procedure 0.128761 0.128761 1 168.139 1.7507 0.1876
## masking 0.123504 0.123504 1 29.767 1.6792 0.2050
## intensity 0.094950 0.094950 1 49.798 1.2910 0.2613
## procedure:masking 0.041057 0.041057 1 100.098 0.5582 0.4567
## procedure:intensity 0.150823 0.150823 1 36.456 2.0506 0.1607
## masking:intensity 0.004736 0.004736 1 60.863 0.0644 0.8005
## procedure:masking:intensity 0.177470 0.177470 1 73.620 2.4129 0.1246## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------------------------
## procedure | 0.01 | [0.00, 1.00]
## masking | 0.05 | [0.00, 1.00]
## intensity | 0.03 | [0.00, 1.00]
## procedure:masking | 5.55e-03 | [0.00, 1.00]
## procedure:intensity | 0.05 | [0.00, 1.00]
## masking:intensity | 1.06e-03 | [0.00, 1.00]
## procedure:masking:intensity | 0.03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].## Warning in term == terms: longer object length is not a multiple of shorter
## object length## Warning in term == names: longer object length is not a multiple of shorter
## object length
Simulated intensity
We ran simulations of acoustic wave propagation and thermal rise to obtain estimates of realized dosage and improved estimates of safety indices. Below you can see the simulated intracranial values for spatial-peak pulse-average intensity (Isppa), thermal rise (\(\Delta\)\(^\circ\)C), and the mechanical index (MI). The values are provided separately for the two applied free-water intensities: 6.35 and 19.05 W/cm2
| mean intensity | SD intensity | mean temperature rise | SD temperature rise | mean mechanical index | SD mechanical index |
|---|---|---|---|---|---|
| 1.788624 | 0.5256269 | 0.0558333 | 0.0359621 | 0.4731703 | 0.0704265 |
| mean intensity | SD intensity | mean temperature rise | SD temperature rise | mean mechanical index | SD mechanical index |
|---|---|---|---|---|---|
| 5.36593 | 1.576897 | 0.2144444 | 0.097474 | 0.8195595 | 0.1219829 |
## [1] "Analyses for lower stimultion intensity"## [1] "Change from baseline"##
## Call:
## lm(formula = p2p_sqrt_baseCor ~ simIntensity, data = cropData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -42.705 -7.866 1.177 13.830 33.776
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -47.769 13.922 -3.431 0.00218 **
## simIntensity 14.862 7.404 2.007 0.05612 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 19.47 on 24 degrees of freedom
## Multiple R-squared: 0.1437, Adjusted R-squared: 0.1081
## F-statistic: 4.029 on 1 and 24 DF, p-value: 0.05612
## [1] "Change from active control"##
## Call:
## lm(formula = p2p_sqrt_vsAC ~ simIntensity, data = cropData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -16.7505 -4.5988 -0.4563 5.7450 24.6293
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.24061 6.23587 0.680 0.503
## simIntensity -0.04407 3.31652 -0.013 0.990
##
## Residual standard error: 8.722 on 24 degrees of freedom
## Multiple R-squared: 7.357e-06, Adjusted R-squared: -0.04166
## F-statistic: 0.0001766 on 1 and 24 DF, p-value: 0.9895
## [1] "Change from sham"##
## Call:
## lm(formula = p2p_sqrt_vsSham ~ simIntensity, data = cropData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -18.640 -9.361 0.977 6.894 25.519
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.853 8.068 0.973 0.340
## simIntensity -2.374 4.291 -0.553 0.585
##
## Residual standard error: 11.28 on 24 degrees of freedom
## Multiple R-squared: 0.01259, Adjusted R-squared: -0.02855
## F-statistic: 0.306 on 1 and 24 DF, p-value: 0.5852
## [1] "Analyses for lower stimultion intensity"## [1] "Change from baseline"##
## Call:
## lm(formula = p2p_sqrt_baseCor ~ simIntensity, data = cropData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -46.518 -14.021 0.513 14.153 41.378
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -46.363 15.781 -2.938 0.00719 **
## simIntensity 4.125 2.798 1.474 0.15340
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 22.07 on 24 degrees of freedom
## Multiple R-squared: 0.08304, Adjusted R-squared: 0.04484
## F-statistic: 2.174 on 1 and 24 DF, p-value: 0.1534
## [1] "Change from active control"##
## Call:
## lm(formula = p2p_sqrt_vsAC ~ simIntensity, data = cropData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -24.906 -7.180 3.592 6.997 24.439
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.805 8.350 0.456 0.653
## simIntensity -1.106 1.480 -0.747 0.462
##
## Residual standard error: 11.68 on 24 degrees of freedom
## Multiple R-squared: 0.02273, Adjusted R-squared: -0.01799
## F-statistic: 0.5581 on 1 and 24 DF, p-value: 0.4623
## [1] "Change from sham"##
## Call:
## lm(formula = p2p_sqrt_vsSham ~ simIntensity, data = cropData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -29.905 -6.605 -2.728 6.384 35.309
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 9.071 9.558 0.949 0.352
## simIntensity -1.867 1.694 -1.102 0.282
##
## Residual standard error: 13.37 on 24 degrees of freedom
## Multiple R-squared: 0.04813, Adjusted R-squared: 0.008472
## F-statistic: 1.214 on 1 and 24 DF, p-value: 0.2815
Blinding to TUS (direct comparison)
Here, we first determine whether masking ultrasonic stimulation effectively blinded participants to whether or not they were receiving TUS. In this analysis, we consider stimulation site (on-target versus active control), masking (unmasked/masked), and intensity, because higher intensities are associated with a louder auditory confound, and thus possibly less effective blinding
## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: yes
## Chisq Df Pr(>Chisq)
## procedure 2.9336 1 0.08676 .
## masking 31.0702 1 2.489e-08 ***
## intensity 4.2932 1 0.03827 *
## procedure:masking 0.5072 1 0.47635
## procedure:intensity 0.6782 1 0.41021
## masking:intensity 0.2528 1 0.61514
## procedure:masking:intensity 0.0177 1 0.89409
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Blinding to TUS (comparison to 50%)
Here, we assess whether discrimintation ability was significantly different from change (i.e., 50%) for each condition.
## [1] "AC_10W_M"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 113.5, p-value = 0.9567
## alternative hypothesis: true location is not equal to 0.5
## [1] "AC_10W_NM"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 190.5, p-value = 0.09578
## alternative hypothesis: true location is not equal to 0.5
## [1] "AC_30W_M"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 159.5, p-value = 0.5043
## alternative hypothesis: true location is not equal to 0.5
## [1] "AC_30W_NM"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 230.5, p-value = 0.003631
## alternative hypothesis: true location is not equal to 0.5
## [1] "Sham"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 146, p-value = 0.4437
## alternative hypothesis: true location is not equal to 0.5
## [1] "TUS_10W_M"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 125, p-value = 0.6914
## alternative hypothesis: true location is not equal to 0.5
## [1] "TUS_10W_NM"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 196, p-value = 0.1732
## alternative hypothesis: true location is not equal to 0.5
## [1] "TUS_30W_M"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 108, p-value = 0.8027
## alternative hypothesis: true location is not equal to 0.5
## [1] "TUS_30W_NM"
##
## Wilcoxon signed rank test with continuity correction
##
## data: participant_level$yes
## V = 186, p-value = 0.01132
## alternative hypothesis: true location is not equal to 0.5
Audible differences
It is possible that, while blinding is effective, audible differences beteween the two stimulation sites remain (i.e., between stimulation of the left-hemispheric hand area and right-hemispheric face area). To test this, participants were presented with pairs of TUS, either both over the same site, or one over each site. These pairs were presented at either 6.35 W/cm2 or 19.05 W/cm2.
## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 same 1 26 28.139 1.51e-05 * 0.177
## 2 intensity 1 26 1.379 2.51e-01 0.004
## 3 same:intensity 1 26 0.856 3.63e-01 0.005
Preperatory learning
Previous research has reported that the amplitude of MEPs is reduced when participants are aware when the TMS pulse will be administered. In the context of our experiments, participants would first need to learn the timing of TMS relative to the onset of TUS. In this scenario, we would expect MEP amplitude on trials with a 500 ms auditory stimulus to begin at amplitudes near baseline MEP amplitude, followed by a reduction in MEP amplitude as participants learn when to expect TMS, and finally more stabilization of MEP amplitude throughout the remainder of the experiment.
Indeed, we see that over the first 50 trials this pattern is observed. To determine whether there is a significant relationship between trial and reduction in MEP amplitude, we split the data into Trial 1-20, and Trial 21-250, to see whether there is a significant decrease of MEP amplitude across initial trials followed by a stabilization.
## [1] "First 20 trials"## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 1.0276 1.0276 1 25.949 16.644 0.0003808 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## [1] "Trial 20-250"## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.014894 0.014894 1 25.942 0.2107 0.65
Experiment III
Part 1
Comparison of sound-sham to baseline
In Part 1, participants underwent 5 conditions: baseline, 1 kHz 500 ms tone, 1 kHz 700 ms tone, 12 kHz 500 ms tone, 12 kHz 700 ms tone. Here, we first test whether each of these conditions significantly reduce MEP amplitude from baseline (i.e., TMS only)
| Baseline |
| 12kHz_long |
| 12kHz_short |
| 1kHz_long |
| 1kHz_short |
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.9989508 0.04819728 15.01644 20.726290 1.842565e-12
## Condition12kHz_long -0.1536070 0.05847780 14.63887 -2.626757 1.934979e-02
## Condition12kHz_short -0.3044756 0.05559294 14.61212 -5.476875 6.999497e-05
## Condition1kHz_long -0.2419641 0.06747553 14.79834 -3.585953 2.754279e-03
## Condition1kHz_short -0.2378596 0.08308604 14.60383 -2.862811 1.210937e-02
Pitch and duration
Next, we directly compare the four different tones to determine whether pitch, duration, or both ultimately affect MEP amplitude.
| 1 kHz |
| 12 kHz |
| short |
| long |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Duration 0.77552 0.77552 1 15.181 7.1159 0.01743 *
## Frequency 0.00265 0.00265 1 14.874 0.0243 0.87825
## Duration:Frequency 0.24353 0.24353 1 14.881 2.2345 0.15586
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## --------------------------------------------------
## Duration | 0.32 | [0.04, 1.00]
## Frequency | 1.63e-03 | [0.00, 1.00]
## Duration:Frequency | 0.13 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Part 2
Ultrasonic stimulation, pitch, and tone duration
In Part 2, the four auditory stimuli were presented again, this time both with and without TUS. Here, we look to see whether there is an effect of ultrasonic stimulation on MEP amplitude over and above presentation of a auditory stimulus alone. We further investigate whether this may be modulated by the tone and/or duration of the auditory stimulus.
| 20W |
| 0W |
| 1 kHz |
| 12 kHz |
| short |
| long |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Procedure 0.04949 0.04949 1 15.423 0.4205 0.52621
## Frequency 0.58098 0.58098 1 14.776 4.9370 0.04235 *
## Duration 0.05173 0.05173 1 15.145 0.4396 0.51729
## Procedure:Frequency 0.02155 0.02155 1 15.159 0.1831 0.67471
## Procedure:Duration 0.49674 0.49674 1 15.077 4.2212 0.05768 .
## Frequency:Duration 0.04537 0.04537 1 14.754 0.3855 0.54414
## Procedure:Frequency:Duration 0.11475 0.11475 1 17.610 0.9751 0.33679
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------------------
## Procedure | 0.03 | [0.00, 1.00]
## Frequency | 0.25 | [0.01, 1.00]
## Duration | 0.03 | [0.00, 1.00]
## Procedure:Frequency | 0.01 | [0.00, 1.00]
## Procedure:Duration | 0.22 | [0.00, 1.00]
## Frequency:Duration | 0.03 | [0.00, 1.00]
## Procedure:Frequency:Duration | 0.05 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
TMS intensity required under different conditions
In Part 1, the TMS intensity (in %MSO) was set to evoke a 1 mV MEP during baseline (i.e., TMS-only). During Part 2, the TMS intensity was set to evoke a 1 mV MEP four different times, once during the presentation of each of the four auditory stimuli. If the required TMS intensity for a 1 mV MEP is higher, that would indicate that MEPs were being attenuated by the auditory stimuli (as indeed observed in Part 1).
## [1] "ANOVA and follow-up"## ANOVA Table (type III tests)
##
## Effect DFn DFd F p p<.05 ges
## 1 Condition 4 60 11.805 3.88e-07 * 0.021##
## Pairwise comparisons using paired t tests
##
## data: MSOlong$MSO and MSOlong$Condition
##
## 12kHz_long 12kHz_short 1kHz_long 1kHz_short
## 12kHz_short 0.49731 - - -
## 1kHz_long 0.47597 0.23900 - -
## 1kHz_short 0.68385 0.42190 0.68232 -
## Baseline 0.00097 0.00188 0.00022 0.00032
##
## P value adjustment method: none## [1] "Mixed model and follow-up"## 1kHz_short 1kHz_long 12kHz_short 12kHz_long
## Baseline 0 0 0 0
## 1kHz_short 1 0 0 0
## 1kHz_long 0 1 0 0
## 12kHz_short 0 0 1 0
## 12kHz_long 0 0 0 1## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 73.6875 2.5557941 15.86455 28.831548 3.938893e-15
## Condition1kHz_short 3.7500 0.6729134 60.00000 5.572782 6.270343e-07
## Condition1kHz_long 3.9375 0.6729134 60.00000 5.851421 2.172582e-07
## Condition12kHz_short 3.2500 0.6729134 60.00000 4.829745 9.819370e-06
## Condition12kHz_long 3.5000 0.6729134 60.00000 5.201263 2.520368e-06## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 171.05 42.762 4 60 11.805 3.882e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## $`emmeans of Condition`
## Condition emmean SE df lower.CL upper.CL
## Baseline 73.7 2.56 15.9 68.3 79.1
## 1kHz_short 77.4 2.56 15.9 72.0 82.9
## 1kHz_long 77.6 2.56 15.9 72.2 83.0
## 12kHz_short 76.9 2.56 15.9 71.5 82.4
## 12kHz_long 77.2 2.56 15.9 71.8 82.6
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
##
## $`pairwise differences of Condition`
## 1 estimate SE df t.ratio p.value
## Baseline - 1kHz_short -3.750 0.673 60 -5.573 <.0001
## Baseline - 1kHz_long -3.938 0.673 60 -5.851 <.0001
## Baseline - 12kHz_short -3.250 0.673 60 -4.830 <.0001
## Baseline - 12kHz_long -3.500 0.673 60 -5.201 <.0001
## 1kHz_short - 1kHz_long -0.188 0.673 60 -0.279 0.7815
## 1kHz_short - 12kHz_short 0.500 0.673 60 0.743 0.4604
## 1kHz_short - 12kHz_long 0.250 0.673 60 0.372 0.7116
## 1kHz_long - 12kHz_short 0.688 0.673 60 1.022 0.3110
## 1kHz_long - 12kHz_long 0.438 0.673 60 0.650 0.5181
## 12kHz_short - 12kHz_long -0.250 0.673 60 -0.372 0.7116
##
## Degrees-of-freedom method: kenward-roger
| Condition | SE | MSO |
|---|---|---|
| Baseline | 2.763933 | 73.6875 |
| 1kHz_short | 2.348703 | 77.4375 |
| 1kHz_long | 2.534553 | 77.6250 |
| 12kHz_short | 2.649636 | 76.9375 |
| 12kHz_long | 2.461739 | 77.1875 |
Â
To verify that the intensities are appropriate, the MEP amplitude should be approximately equal across all the four applied intensities.
| Condition | SE | p2p_sqrt |
|---|---|---|
| 1kHz_short_0W | 0.0678416 | 0.8914426 |
| 1kHz_long_0W | 0.0720263 | 0.9032426 |
| 12kHz_short_0W | 0.0683276 | 1.0047251 |
| 12kHz_long_0W | 0.0602894 | 0.9288223 |
| Baseline | 0.0476694 | 0.9995522 |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 0.60996 0.15249 4 14.94 1.2815 0.3211
While the MEP amplitudes do not significantly differ from each other, they do differ from one another up to 13%.
In Part I, there was a baseline condition. In Part 2, however, there was not. Nevertheless, we would like to know whether motor inhibition took place (as compared to baseline) for the conditions in Part 2. Therefore, we calculate the Adjusted MEP Amplitude. This measure will take into account not only the different TMS intensities applied, but also account for the differences in the TMS intensity’s ability to evoke a 1 mV MEP.
Below, you can see for each condition where TMS intensity was set to evoke a 1 mV amplitude, the factor required to actually have evoked exactly a 1 mV amplitude.
| Condition | ParticipantCode | SE | p2p_sqrt | normalizeFactor |
|---|---|---|---|---|
| 1kHz_short_0W | S1 | 0.0780526 | 0.8470204 | 1.1806091 |
| 1kHz_short_0W | S10 | 0.0634664 | 1.5055051 | 0.6642289 |
| 1kHz_short_0W | S11 | 0.0594106 | 0.9000684 | 1.1110267 |
| 1kHz_short_0W | S12 | 0.0587072 | 0.3242531 | 3.0840109 |
| 1kHz_short_0W | S13 | 0.0921704 | 1.0257319 | 0.9749136 |
| 1kHz_short_0W | S14 | 0.0323222 | 0.8044612 | 1.2430681 |
| 1kHz_short_0W | S15 | 0.0965106 | 0.7300764 | 1.3697197 |
| 1kHz_short_0W | S16 | 0.0623067 | 0.7633442 | 1.3100250 |
| 1kHz_short_0W | S2 | 0.0648627 | 0.9620159 | 1.0394838 |
| 1kHz_short_0W | S3 | 0.0722470 | 1.2341461 | 0.8102768 |
| 1kHz_short_0W | S4 | 0.0855365 | 0.9796650 | 1.0207571 |
| 1kHz_short_0W | S5 | 0.0621767 | 0.8960877 | 1.1159622 |
| 1kHz_short_0W | S6 | 0.1073580 | 1.1672370 | 0.8567241 |
| 1kHz_short_0W | S7 | 0.0638458 | 0.8201140 | 1.2193427 |
| 1kHz_short_0W | S8 | 0.1281685 | 0.5832421 | 1.7145538 |
| 1kHz_short_0W | S9 | 0.1051449 | 0.7201136 | 1.3886699 |
| 1kHz_long_0W | S1 | 0.0594045 | 0.7053159 | 1.4178045 |
| 1kHz_long_0W | S10 | 0.0773279 | 1.3568678 | 0.7369915 |
| 1kHz_long_0W | S11 | 0.0728205 | 0.9197869 | 1.0872084 |
| 1kHz_long_0W | S12 | 0.0668062 | 0.7086205 | 1.4111925 |
| 1kHz_long_0W | S13 | 0.1454348 | 1.4254278 | 0.7015438 |
| 1kHz_long_0W | S14 | 0.0776822 | 0.7340906 | 1.3622297 |
| 1kHz_long_0W | S15 | 0.0824091 | 0.6660776 | 1.5013266 |
| 1kHz_long_0W | S16 | 0.0397827 | 0.5778280 | 1.7306189 |
| 1kHz_long_0W | S2 | 0.0602011 | 0.7079422 | 1.4125446 |
| 1kHz_long_0W | S3 | 0.0730157 | 1.3576332 | 0.7365760 |
| 1kHz_long_0W | S4 | 0.0705659 | 0.6064539 | 1.6489299 |
| 1kHz_long_0W | S5 | 0.1030069 | 1.0107925 | 0.9893227 |
| 1kHz_long_0W | S6 | 0.0802677 | 0.9777180 | 1.0227898 |
| 1kHz_long_0W | S7 | 0.0593773 | 0.6292481 | 1.5891983 |
| 1kHz_long_0W | S8 | 0.1453109 | 1.1744592 | 0.8514557 |
| 1kHz_long_0W | S9 | 0.0871747 | 0.8936195 | 1.1190445 |
| 12kHz_short_0W | S1 | 0.0661769 | 0.9456057 | 1.0575233 |
| 12kHz_short_0W | S10 | 0.0771140 | 1.3270829 | 0.7535324 |
| 12kHz_short_0W | S11 | 0.0722597 | 1.0749024 | 0.9303170 |
| 12kHz_short_0W | S12 | 0.0813503 | 0.8454025 | 1.1828685 |
| 12kHz_short_0W | S13 | 0.1283424 | 0.8183051 | 1.2220381 |
| 12kHz_short_0W | S14 | 0.0707396 | 0.8684779 | 1.1514398 |
| 12kHz_short_0W | S15 | 0.0961953 | 0.7102773 | 1.4079009 |
| 12kHz_short_0W | S16 | 0.0502684 | 0.5256514 | 1.9024014 |
| 12kHz_short_0W | S2 | 0.0830281 | 0.8648062 | 1.1563284 |
| 12kHz_short_0W | S3 | 0.0837238 | 1.3731674 | 0.7282434 |
| 12kHz_short_0W | S4 | 0.0954151 | 0.7778999 | 1.2855124 |
| 12kHz_short_0W | S5 | 0.0868036 | 1.2888419 | 0.7758904 |
| 12kHz_short_0W | S6 | 0.1262734 | 1.0697581 | 0.9347908 |
| 12kHz_short_0W | S7 | 0.0981261 | 1.1125775 | 0.8988138 |
| 12kHz_short_0W | S8 | 0.1029919 | 1.5546259 | 0.6432416 |
| 12kHz_short_0W | S9 | 0.0842689 | 0.9182201 | 1.0890635 |
| 12kHz_long_0W | S1 | 0.0736127 | 0.7510728 | 1.3314288 |
| 12kHz_long_0W | S10 | 0.0688922 | 1.2703267 | 0.7871991 |
| 12kHz_long_0W | S11 | 0.0441020 | 0.9313775 | 1.0736784 |
| 12kHz_long_0W | S12 | 0.1055087 | 0.9458055 | 1.0572998 |
| 12kHz_long_0W | S13 | 0.1372278 | 0.9032992 | 1.1070530 |
| 12kHz_long_0W | S14 | 0.0836243 | 0.7973843 | 1.2541004 |
| 12kHz_long_0W | S15 | 0.0439669 | 0.6359112 | 1.5725466 |
| 12kHz_long_0W | S16 | 0.0361035 | 0.6587108 | 1.5181168 |
| 12kHz_long_0W | S2 | 0.0602011 | 0.7079422 | 1.4125446 |
| 12kHz_long_0W | S3 | 0.0569948 | 1.4871731 | 0.6724167 |
| 12kHz_long_0W | S4 | 0.1176358 | 0.8119685 | 1.2315748 |
| 12kHz_long_0W | S5 | 0.0881689 | 1.3171626 | 0.7592077 |
| 12kHz_long_0W | S6 | 0.1587037 | 1.0126376 | 0.9875201 |
| 12kHz_long_0W | S7 | 0.0723476 | 0.8519831 | 1.1737322 |
| 12kHz_long_0W | S8 | 0.1610778 | 0.9443384 | 1.0589424 |
| 12kHz_long_0W | S9 | 0.0847524 | 0.8340639 | 1.1989490 |
| Baseline | S1 | 0.0831449 | 1.0870603 | 0.9199122 |
| Baseline | S10 | 0.0641499 | 1.1361589 | 0.8801586 |
| Baseline | S11 | 0.1428946 | 1.2554183 | 0.7965473 |
| Baseline | S12 | 0.1016682 | 0.9854720 | 1.0147422 |
| Baseline | S13 | 0.1426936 | 1.1357613 | 0.8804667 |
| Baseline | S14 | 0.0641702 | 0.9014603 | 1.1093112 |
| Baseline | S15 | 0.0644333 | 0.6655947 | 1.5024158 |
| Baseline | S16 | 0.1319401 | 1.1845919 | 0.8441726 |
| Baseline | S2 | 0.0720395 | 0.8031923 | 1.2450319 |
| Baseline | S3 | 0.1006844 | 1.1022015 | 0.9072752 |
| Baseline | S4 | 0.0570554 | 0.7626954 | 1.3111394 |
| Baseline | S5 | 0.1099815 | 1.0686523 | 0.9357581 |
| Baseline | S6 | 0.1480828 | 1.1743366 | 0.8515446 |
| Baseline | S7 | 0.0810162 | 0.8716353 | 1.1472688 |
| Baseline | S8 | 0.1819932 | 1.1666718 | 0.8571391 |
| Baseline | S9 | 0.0977517 | 0.6919320 | 1.4452286 |
Now that the MEP amplitudes have been adjusted to on average 1 mV per applied intensity, we can correct MEP amplitudes for different applied intensities.
This is how MEP amplitudes are transformed for the 5 conditions where TMS was intended to evoke a 1mV MEP:
| Condition | SE | p2p_sqrt |
|---|---|---|
| 1kHz_short_0W | 0 | 1 |
| 1kHz_long_0W | 0 | 1 |
| 12kHz_short_0W | 0 | 1 |
| 12kHz_long_0W | 0 | 1 |
| Baseline | 0 | 1 |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 2.2382e-28 5.5955e-29 4 1162 0 1
Now a correction factor is applied to equalize the TMS intensities to baseline
| Condition | SE | p2p_sqrt |
|---|---|---|
| 1kHz_short_0W | 0.0110326 | 0.9491366 |
| 1kHz_long_0W | 0.0104425 | 0.9477691 |
| 12kHz_short_0W | 0.0110941 | 0.9571101 |
| 12kHz_long_0W | 0.0112308 | 0.9527400 |
| Baseline | 0.0000000 | 1.0000000 |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Condition 0.45676 0.11419 4 1162 0.8907 0.4688
First, we test whether each condition in Part 2 reduces from baseline
## 12kHz_long_0W 12kHz_long_20W 12kHz_short_0W 12kHz_short_20W
## Baseline 0 0 0 0
## 12kHz_long_0W 1 0 0 0
## 12kHz_long_20W 0 1 0 0
## 12kHz_short_0W 0 0 1 0
## 12kHz_short_20W 0 0 0 1
## 1kHz_long_0W 0 0 0 0
## 1kHz_long_20W 0 0 0 0
## 1kHz_short_0W 0 0 0 0
## 1kHz_short_20W 0 0 0 0
## 1kHz_long_0W 1kHz_long_20W 1kHz_short_0W 1kHz_short_20W
## Baseline 0 0 0 0
## 12kHz_long_0W 0 0 0 0
## 12kHz_long_20W 0 0 0 0
## 12kHz_short_0W 0 0 0 0
## 12kHz_short_20W 0 0 0 0
## 1kHz_long_0W 1 0 0 0
## 1kHz_long_20W 0 1 0 0
## 1kHz_short_0W 0 0 1 0
## 1kHz_short_20W 0 0 0 1## Estimate Std. Error df t value
## (Intercept) 0.99828001 0.04784225 15.03389 20.8660762
## Condition12kHz_long_0W -0.11772795 0.04898765 15.67376 -2.4032169
## Condition12kHz_long_20W -0.04941609 0.05242160 15.24092 -0.9426666
## Condition12kHz_short_0W -0.03957859 0.05936217 15.01127 -0.6667308
## Condition12kHz_short_20W -0.11062596 0.06047875 14.93860 -1.8291709
## Condition1kHz_long_0W -0.14194642 0.05665517 15.03911 -2.5054451
## Condition1kHz_long_20W -0.13572020 0.06286661 15.09517 -2.1588598
## Condition1kHz_short_0W -0.15277763 0.06774374 14.90185 -2.2552286
## Condition1kHz_short_20W -0.21646148 0.07038850 14.79849 -3.0752392
## Pr(>|t|)
## (Intercept) 1.634993e-12
## Condition12kHz_long_0W 2.901591e-02
## Condition12kHz_long_20W 3.605553e-01
## Condition12kHz_short_0W 5.150620e-01
## Condition12kHz_short_20W 8.740488e-02
## Condition1kHz_long_0W 2.420917e-02
## Condition1kHz_long_20W 4.736167e-02
## Condition1kHz_short_0W 3.959608e-02
## Condition1kHz_short_20W 7.796576e-03## $`emmeans of Condition`
## Condition emmean SE df lower.CL upper.CL
## Baseline 0.998 0.0479 15.0 0.896 1.100
## 12kHz_long_0W 0.881 0.0528 15.0 0.768 0.993
## 12kHz_long_20W 0.949 0.0683 15.0 0.803 1.095
## 12kHz_short_0W 0.959 0.0649 15.0 0.820 1.097
## 12kHz_short_20W 0.888 0.0711 15.0 0.736 1.039
## 1kHz_long_0W 0.856 0.0683 15.0 0.711 1.002
## 1kHz_long_20W 0.863 0.0667 14.8 0.720 1.005
## 1kHz_short_0W 0.846 0.0657 15.0 0.706 0.985
## 1kHz_short_20W 0.782 0.0615 15.0 0.651 0.913
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
##
## $`pairwise differences of Condition`
## 1 estimate SE df t.ratio p.value
## Baseline - 12kHz_long_0W 0.11773 0.0490 15.0 2.401 0.0298
## Baseline - 12kHz_long_20W 0.04942 0.0525 15.0 0.942 0.3611
## Baseline - 12kHz_short_0W 0.03958 0.0594 15.0 0.666 0.5153
## Baseline - 12kHz_short_20W 0.11063 0.0605 15.0 1.828 0.0875
## Baseline - 1kHz_long_0W 0.14195 0.0567 15.0 2.504 0.0243
## Baseline - 1kHz_long_20W 0.13572 0.0638 14.8 2.129 0.0505
## Baseline - 1kHz_short_0W 0.15278 0.0678 15.0 2.255 0.0395
## Baseline - 1kHz_short_20W 0.21646 0.0704 15.0 3.074 0.0077
## 12kHz_long_0W - 12kHz_long_20W -0.06831 0.0514 15.0 -1.328 0.2039
## 12kHz_long_0W - 12kHz_short_0W -0.07815 0.0448 15.0 -1.745 0.1015
## 12kHz_long_0W - 12kHz_short_20W -0.00710 0.0663 15.0 -0.107 0.9162
## 12kHz_long_0W - 1kHz_long_0W 0.02422 0.0463 15.0 0.523 0.6085
## 12kHz_long_0W - 1kHz_long_20W 0.01799 0.0507 14.5 0.355 0.7276
## 12kHz_long_0W - 1kHz_short_0W 0.03505 0.0619 15.0 0.566 0.5798
## 12kHz_long_0W - 1kHz_short_20W 0.09873 0.0651 15.0 1.517 0.1501
## 12kHz_long_20W - 12kHz_short_0W -0.00984 0.0502 15.0 -0.196 0.8473
## 12kHz_long_20W - 12kHz_short_20W 0.06121 0.0521 15.0 1.175 0.2581
## 12kHz_long_20W - 1kHz_long_0W 0.09253 0.0459 15.0 2.015 0.0621
## 12kHz_long_20W - 1kHz_long_20W 0.08630 0.0594 14.7 1.453 0.1671
## 12kHz_long_20W - 1kHz_short_0W 0.10336 0.0620 15.0 1.668 0.1160
## 12kHz_long_20W - 1kHz_short_20W 0.16705 0.0754 15.0 2.216 0.0425
## 12kHz_short_0W - 12kHz_short_20W 0.07105 0.0587 15.0 1.210 0.2450
## 12kHz_short_0W - 1kHz_long_0W 0.10237 0.0566 15.0 1.808 0.0907
## 12kHz_short_0W - 1kHz_long_20W 0.09614 0.0524 14.6 1.834 0.0872
## 12kHz_short_0W - 1kHz_short_0W 0.11320 0.0778 15.0 1.455 0.1662
## 12kHz_short_0W - 1kHz_short_20W 0.17688 0.0787 15.0 2.249 0.0400
## 12kHz_short_20W - 1kHz_long_0W 0.03132 0.0595 15.0 0.526 0.6063
## 12kHz_short_20W - 1kHz_long_20W 0.02509 0.0650 14.8 0.386 0.7051
## 12kHz_short_20W - 1kHz_short_0W 0.04215 0.0774 15.0 0.545 0.5940
## 12kHz_short_20W - 1kHz_short_20W 0.10584 0.0738 15.0 1.435 0.1719
## 1kHz_long_0W - 1kHz_long_20W -0.00623 0.0495 14.5 -0.126 0.9016
## 1kHz_long_0W - 1kHz_short_0W 0.01083 0.0641 15.0 0.169 0.8680
## 1kHz_long_0W - 1kHz_short_20W 0.07452 0.0753 15.0 0.989 0.3384
## 1kHz_long_20W - 1kHz_short_0W 0.01706 0.0713 14.9 0.239 0.8143
## 1kHz_long_20W - 1kHz_short_20W 0.08074 0.0764 14.9 1.057 0.3075
## 1kHz_short_0W - 1kHz_short_20W 0.06368 0.0482 15.0 1.321 0.2063
##
## Degrees-of-freedom method: kenward-roger## `summarise()` has grouped output by 'ParticipantCode'. You can override using
## the `.groups` argument.## # A tibble: 143 × 6
## # Groups: ParticipantCode [16]
## ParticipantCode Condition p2p_sqrt p2p_sqrt_MSOcor SE AdjustedMEP
## <chr> <fct> <dbl> <dbl> <dbl> <dbl>
## 1 S1 Baseline 1.09 1.09 0.0765 1
## 2 S1 12kHz_long_0W 0.751 0.716 0.0935 0.954
## 3 S1 12kHz_long_20W 1.11 1.06 0.0884 1.41
## 4 S1 12kHz_short_0W 0.946 0.946 0.0700 1
## 5 S1 12kHz_short_20W 0.913 0.913 0.121 0.965
## 6 S1 1kHz_long_0W 0.705 0.653 0.0779 0.925
## 7 S1 1kHz_long_20W 0.895 0.829 0.150 1.17
## 8 S1 1kHz_short_0W 0.847 0.772 0.0840 0.912
## 9 S1 1kHz_short_20W 0.846 0.771 0.0727 0.910
## 10 S10 Baseline 1.14 1.14 0.0565 1
## # … with 133 more rows## [1] "For all comparisons but 1kHz_long_20W (because one missing)"##
## Pairwise comparisons using paired t tests
##
## data: cropData$p2p_sqrt_MSOcor and cropData$Condition
##
## Baseline 12kHz_long_0W 12kHz_long_20W 12kHz_short_0W
## 12kHz_long_0W 0.0273 - - -
## 12kHz_long_20W 0.3420 0.2048 - -
## 12kHz_short_0W 0.5093 0.0984 0.8241 -
## 12kHz_short_20W 0.0832 0.9198 0.2527 0.2349
## 1kHz_long_0W 0.0227 0.5937 0.0593 0.0842
## 1kHz_short_0W 0.0372 0.5720 0.1138 0.1618
## 1kHz_short_20W 0.0076 0.1486 0.0449 0.0406
## 12kHz_short_20W 1kHz_long_0W 1kHz_short_0W
## 12kHz_long_0W - - -
## 12kHz_long_20W - - -
## 12kHz_short_0W - - -
## 12kHz_short_20W - - -
## 1kHz_long_0W 0.5989 - -
## 1kHz_short_0W 0.5924 0.8713 -
## 1kHz_short_20W 0.1749 0.3475 0.2106
##
## P value adjustment method: none## [1] "For 1kHz_long_20S comparison"##
## Paired t-test
##
## data: p2p_sqrt by Condition
## t = 1.3565, df = 14, p-value = 0.1964
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.05656422 0.25124492
## sample estimates:
## mean of the differences
## 0.09734035
Finally, we do pairwise comparisons between each auditory stimulus, with and without TUS.
## [1] "Direct comparison 1 kHz Short: sham versus TUS"##
## Paired t-test
##
## data: p2p_sqrt by Condition
## t = 1.347, df = 15, p-value = 0.198
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.0392928 0.1742229
## sample estimates:
## mean of the differences
## 0.06746504## [1] "Direct comparison 1 kHz Long: sham versus TUS"##
## Paired t-test
##
## data: p2p_sqrt by Condition
## t = -0.10501, df = 14, p-value = 0.9179
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.1161602 0.1053161
## sample estimates:
## mean of the differences
## -0.005422042## [1] "Direct comparison 12 kHz Short: sham versus TUS"##
## Paired t-test
##
## data: p2p_sqrt by Condition
## t = 1.2658, df = 15, p-value = 0.2249
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.05283143 0.20734042
## sample estimates:
## mean of the differences
## 0.07725449## [1] "Direct comparison 12 kHz Long: sham versus TUS"##
## Paired t-test
##
## data: p2p_sqrt by Condition
## t = -1.2931, df = 15, p-value = 0.2155
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.18276462 0.04474396
## sample estimates:
## mean of the differences
## -0.06901033Preperatory learning
Finally, like in Experiment I and II, we look into whether there is evidence for preperatory learning.
## [1] "For the first block with sound (regardless of duration)"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.10329 0.10329 1 14.718 1.3402 0.2654
## [1] "For the first block with 500 ms sound"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.0192 0.0192 1 2.038 0.1745 0.716
## [1] "For the first block with 700 ms sound"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.10329 0.10329 1 14.718 1.3402 0.2654
## [1] "Trial number per block in subsequent blocks:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 0.16056 0.16056 1 14.602 1.2506 0.2815
## [1] "Trial number in total:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.12648 0.12648 1 15.476 0.9788 0.3377
Experiment IV
Stimulation Intensity and Masking (two variants)
## [1] "For 3-level intensity:"| low |
| medium |
| high |
| unmasked |
| masked |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Masking 14.0884 14.0884 1 4053.1 70.815 < 2.2e-16 ***
## Intensity 8.3221 4.1611 2 4053.0 20.916 9.185e-10 ***
## Masking:Intensity 5.2178 2.6089 2 4053.0 13.114 2.105e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -------------------------------------------------
## Masking | 0.02 | [0.01, 1.00]
## Intensity | 0.01 | [0.01, 1.00]
## Masking:Intensity | 6.43e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "For 6-level intensity:"| low |
| medium |
| high |
| unmasked |
| masked |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Masking 54169 54169 1 4048.8 52.2984 5.673e-13 ***
## Intensity6 23593 4719 5 4048.2 4.5557 0.0003813 ***
## Masking:Intensity6 7602 1520 5 4048.4 1.4680 0.1968304
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## --------------------------------------------------
## Masking | 0.01 | [0.01, 1.00]
## Intensity6 | 5.60e-03 | [0.00, 1.00]
## Masking:Intensity6 | 1.81e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
There is a significant interaction between masking and intensity. The analysis will not be split by masking.
## [1] "Testing the interaction for 3-level intensity:"## [1] "Analysis of Intensity (for unmasked conditions"| low |
| medium |
| high |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Intensity 4213.1 2106.5 2 1974.7 2.1724 0.1142## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Intensity | 2.20e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Analysis of Intensity (for masked conditions"| low |
| medium |
| high |
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Intensity 6587.8 3293.9 2 2070.4 3.0298 0.04854 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Intensity | 2.92e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Comparison to baseline
For 3 level
## [1] "For 3-level intensity:"## [1] "Baseline comparison for Audio_deep_30_3"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 0.56591 0.56591 1 11.287 4.166 0.06534 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.27 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_3"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 4.2209 4.2209 1 10.692 26.812 0.0003339 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.71 | [0.40, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_6"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 2.1099 2.1099 1 11.168 11.283 0.006247 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.50 | [0.13, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_6"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 5.8219 5.8219 1 11.114 37.008 7.584e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.77 | [0.50, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_9"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 2.4707 2.4707 1 11.224 13.389 0.00364 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.54 | [0.17, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_9"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 2.9368 2.9368 1 11.03 21.155 0.0007602 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.66 | [0.31, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
For 6 level
## [1] "For 6-level intensity:"## [1] "Baseline comparison for Audio_deep_30_2"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 0.41531 0.41531 1 11.258 3.6073 0.08344 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.24 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_2"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 4.1215 4.1215 1 11.194 26.399 0.0003066 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.70 | [0.39, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_3"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 0.22797 0.22797 1 11.492 1.7812 0.2078
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.13 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_3"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 2.3108 2.3108 1 9.9238 19.013 0.001446 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.66 | [0.29, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_5"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 1.2488 1.2488 1 11.1 7.849 0.01709 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.41 | [0.06, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_5"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 2.7739 2.7739 1 11.293 21.188 0.0007109 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.65 | [0.31, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_6"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 0.84875 0.84875 1 10.892 6.396 0.02821 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.37 | [0.03, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_6"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 1.5753 1.5753 1 9.9256 13.826 0.004038 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.58 | [0.19, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_8"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 1.6528 1.6528 1 11.184 11.913 0.005287 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.52 | [0.14, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_8"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 3.6029 3.6029 1 9.8513 27.009 0.0004231 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.73 | [0.41, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for Audio_deep_30_9"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 0.92197 0.92197 1 11.319 6.4167 0.02729 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.36 | [0.03, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## [1] "Baseline comparison for NoAudio_deep_30_9"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Baseline 1.136 1.136 1 11.132 9.5289 0.01021 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------
## Baseline | 0.46 | [0.09, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
Preperatory learning
## [1] "For the first block with TUS (no masking)"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 0.014614 0.014614 1 11.526 0.1404 0.7147
## [1] "Trial number per block:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 0.2004 0.2004 1 11.252 1.1022 0.3158
## [1] "Trial number in total:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 0.20027 0.20027 1 11.254 1.102 0.3159
Baseline corrected APB as dependent variable
## [1] "For the first block with TUS (no masking)"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 787.02 787.02 1 11.211 1.0112 0.3358
## [1] "Trial number per block:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## trialblock 317.72 317.72 1 10.962 0.3355 0.5741
## [1] "Trial number in total:"
## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Trial 327.68 327.68 1 10.963 0.346 0.5683